Hello.

They can definitely be done in house and it's not difficult. MS detectors will be more tricky and will require more expensive kits. The easiest way to build a procedure and specs if you haven't done it before is by going through the paperwork the vendors gave you when they installed the machines for the first time and also through the annual reports from prior years. That paperwork will have all tests they did and the specifications they needed to meet. Additionally, the purchase agreement from both Waters and Agilent usually contains detailed description for the OQ they were to perform.

Waters sells maintenance kits (valves, seal packs, tubes, lamps, etc) and calibration standards (uracil and caffeine for peak area, height, retention time, and resolution) with the required HPLC column. Typical parameters you do are flow rate accuracy, column oven temperature range, wavelength accuracy, detector noise, and peak retention time/area/height precision, injection, response linearity for UV detectors, carryover check, and gradient mix accuracy to test different lines.

Onto your questions:

Has anyone done it?

Yes, I have done HPLC with UV and PDA and GC-HS with FID. I chose to let Waters to their MS detectors. I also used to do some equipment administration in our company.

Is it easy/straightforward/expensive?

It's easy and straightforward but you need to review in details the manual and parameters for each separate instrument. For example, a PDA and a UV detector will have different noise criteria, and different brands/models of HPLC will have different accuracy of flow rate. Additionally, if you are controlled environment that gets inspected by authorities, you need to have clear and detailed SOP about how to perform all this, what specs are to be met, how often to perform it, and who is qualified to perform the work. If your environment is not heavily controlled and your results aren't used for human products/tests, then it's much simpler.

As for the cost, it's cheaper to do it by yourself. By much? Not really. If you have low number of units and you can do them in a week or so at the beginning or end of the year and then it's easy and cheaper. If you have several machines, to me it's not worth it unless you have a dedicated person or a metrologist to take care of all machines. The PM kits and UV lamps cost decent money and if you need to do several machines by yourself, that will take a big chunk of time (which if you calculate it as hourly rate will come high). I work in the pharmaceutical business when things go very fast and priories change weekly. We have many units of instruments and it takes time to do all of them. Collecting, interpreting, and reviewing data is my main priority, therefore as long as the company is not burdened by the cost, I choose to have the manufactures do the annual service and qualification/checks.

Mr. Solntsev,

The answer to your question depends on the activity of your Company.

If, the analytical instrumentation is used to scientific purposes, then you are completely able to calibrate your instrumentation alone; thus, operating at the so-called ''academic accuracy'' of the measurements.

However, if your Company provides certified analyses according to ISOs, for example, then you should use the services provided by Agilent, in order to operate at the legislation accuracy. They ensure an assessment of the interlaboratory variability of your data. In other words, in this latter case you should not calibrate your instrumentation alone.

(The mass spectrometric analyses have property of physical evidence in front of the Legislation Authorities.)

Hi

You are right

It is not mandatory to go to external calibration unless it is required, being a small organisation cost control is vital. You can calibrate in house and maintain records of calibration. Go through the operational manual of HPLC supplied along with instrument at the time of purchasing and follow for calibration. Buying kit is ideal or you can buy reference standard samples available with authenticated certificates. These reference samples are cheaper than kits.

Good luck

I’m fortunate enough to have had no direct contact with this business for about 15 years. However, from occasional browsing I gather that some wasteful and irrational protocols and procedures are still in force.

The present calibration / PQ /OQ system was introduced a couple of decades ago to ensure high laboratory standards, but it seems to have engendered a management / subcontracting / inspecting system with blind checking of checkboxes, that deprives the scientists who use the instruments of expertise and responsibility. There is no notion of what is critical and what is less so, nor of the idea that more and better performance data can be obtained in-use, from daily system suitability and calibration procedures.

I’ve seen no attempt to distinguish critical parameters (for example temperatures for physical constants, stability testing chambers, dissolution testing of solid dosage forms) from ones that are not (for example all the temperatures that are set in GC, LC and MS).

Chromatography is possibly the field where useless practices are most evident. HPLC column packings are composite proprietary materials that can vary and deteriorate ; GC stationary phases are better characterised, though relevant dimensions (phase thickness, column i.d. and length) can vary. Consequently, given variations in chromatographic parameters like retention time and peak height and width, and the necessary adjustments during system suitability testing, quantitative analyses are invariably calibrated with respect to the respective reference substances. Therefore, few instrumental parameters are critical regarding accuracy. For example, since the absorbance measured by an LC/UV detector doesn’t need to be terribly accurate, you should not need a special certified reference substances to check that. As I’ve said, relevant data (repeatability, detector linearity…) is obtained in-use, perhaps during a method validation exercise. Detector noise is evaluated during every run.

People pay good money to certify a temperature probe designed to fit into the GC injector. In our lab we paid 300 euros to certify with respect to national standards the pressure gauge with needle that was used to « qualify » the carrier gas pressure, which you are allowed (is that still true ?) to adjust anyway. One sub-contractor had a procedure for UV detector wavelength that involved removing the flow cell ; the manufacturer’s manual stated that you have to check the wavelenth whenever the cell is removed, but the tastefully-designed qualifiaction certificate pleased company management and the inspectorate. By the way, the best time to change a UV lamp is when system suitability shows it will soon become necessary.

Great emphasis is placed on testing balances using a single expensive reference weight, whereas the accuracy is not critical for analyses where you weigh your reference substance on the same balance. What is critical is balance linearity. No-one knew how to verify that with an electronic balance, though the subject was considered important when each balance had its own set of weights. Concerning electronic balances, can one be certain that none of the bits of the analogue-digital converter (which may have 24 bits) has become latched high or low? That happened to us with some LC-UV detectors during a period of electromagnetic interference; it was noticeable only because chromatograms are continuous recordings. I never saw a SOP that tried to distinguish analytical balances used for comparative purposes, and balances used for determining the mass of buffer ingredients and dosage forms or for weighing ingredients during manufacturing.

To conclude this rant, I propose it would be time for those responsible for standards and regulations to take another look at principles and practices in the calibration and qualification scene.

Is this for formal validation or regulatory use? *The answer is really important. Those kits you mention are just pre-packaged supplies, not the SOPs, Policies and documents needed to carry out a qualification (training is not included).

• If it is for formal validation to meet regulatory requirements then, the requirements are rather strict and time consuming to get everything needed ready (and also prove the persons authoring, reviewing and doing the actual work are qualified to do so).
• If this is just for training purposes and you were not actually interested in formal qualifications (IQ, OQ), but instead only preventative maintenance (PM), then that is less time consuming (and you should still use SOPs and Policies). *My company has been doing this type of work professionally for decades and it is a LOT of work when you do not already have the dedicated SOPs in place.

If (1) you have someone with a great deal of experience in both chromatography AND the systems you use and (2) you have all of the required SOP's needed that you created in-house, which are specific to your instrument modules and (3) all of the standards, test equipment and columns specified in your SOP's to carry out each test (such as columns, test stds, tubing, calibrated probes etc), then yes, you can do this in-house. Most clients hire the manufacture to do so because they already have all of the required SOPs, training certificates and documentation needed to issue the reports. **Take a look at the documents supplied by the original manufacturer after hiring them to do it the first time and you will see how involved it can be.

Note: For a typical HPLC instrument IQ, OQ or even a PV, the amount of paperwork alone is huge and takes months to generate for most modules and test methods (i.e. new Policy forms, SOP's and training / equipment logs). Once initially generated, then the whole process only takes 1 or 2 days to complete per instrument for an IQ or OQ. GC-MSD is much simpler to initially create, but you still need the SOP's and Policy documents in place.

BTW: None of those qualification documents verifies that you know what you are doing or that any methods run on that instrument will produce accurate data. The purpose of all this qualification documentation is only to verify that the instrument meets the manufacturer's original specifications, that is it. It provides an initial bench-mark, a way to insure that the equipment can pass a test, but tells nothing about how it is used or the validity of the data generated. Much as 'Method Validation' also does not verify that the results obtained are valid either. IMO: Invest your money in more training, not doing IQ/OQs in-house. Hire the manufacturer to perform the IQ/OQ's. Training is what makes the real difference in running valid methods, maintaining the instrument(s) and generating accurate data. Professional training in method development, operation and troubleshooting are what gets skipped at most companies.